Cycle to Perth

I cycled to Perth yesterday, mainly as a way to stop thinking about things. I didn’t really think there would be any OSM mapping to be done on such a popular route, and there wouldn’t have been if it weren’t for the roadworks along the freeway

which forced me to cross over the overpass and ride through the streets of South Perth. The official detour ended and I could have crossed back, but I carried on… through non-cyclepaths and large detours around intersections.

I looked at the map, and it did actually look like there was a whole path along the golf course that wasn’t on the map, but actually it was added a month ago and OSMand hasn’t updated yet.

But it was interesting to see the new timber shingles being put on the Old Mill.

2019-08-17 The Old Mill, roof renovations 1.jpg
File:2019-08-17_The_Old_Mill,_roof_renovations_1.jpg (17 August 2019, 13:47:05) by Samwilson, CC-BY-SA-4.0.

English: New shingles being installed on the Old Mill, South Perth.

DDD Perth 2019

On Saturday I went to my second DDD Perth conference. This is an annual one-day event that seems to bill itself as being something different from the usual tech conference and to be more accessible to people who might not normally go to these things. Which, I reckon, is pretty great; it feels like a more friendly place (even though I am the sort of person who usually goes to these sorts of things). There is something of a corporate vibe, though, which sort of mars the ‘community’ aspect — although that’s where the low price of the tickets comes from, so I shouldn’t complain!

The welcome to country was done by Nick Abraham, who gave an interesting talk. I liked the lists that he rattled off the names of families he’s part of and places they’re from.

All of the presentations that I went to were interesting, but the two that have stuck with me were about CSS Grid (Amy Kapernick), and WebAuthn (Ben Lowry). I want to experiment with both in my personal projects. Don’t suppose I’ll do so, though. (Too many fun things to explore; not enough time.)

I’m very glad that the Perth developer community is able to support this sort of event these days.

Backing up (my) Commons files

I’m experimenting with an idea of treating the Commons copies of my photos as the ‘master’ copy, and not keeping them online anywhere else (e.g. Flickr). This involves uploading to Commons and then keeping a local copy in sync — because I don’t want to lose any photos if they get deleted from Commons.

I’m using Digikam locally, and have two collections configured: one scratch-pad one, for sorting out photos that are just off the camera; and one backup one, which lets me browse photos I’ve got on Commons.

I download from Commons with the following backup.sh script, which goes through all of my contributions and exports XML for every page I’ve worked on, and every file for which I’m the first author (i.e. I uploaded it).

#!/bin/bash
BACKUP_DIR=$(cd "$(dirname $0)"; pwd -P)
mwcli export:contribs \
	--config="$BACKUP_DIR/config.yml" \
	--wiki=commons \
	--user=samwilson \
	--dest="$BACKUP_DIR" \
	--only-author

The mwcli script is at github.com/samwilson/mwcli

The reason I want the Commons copy to be canonical is that it makes for centralised metadata, a single place to edit and add links to related material. It’s annoying to have to keep metadata in sync between Commons, Flickr, and possibly a local copy of things too.

Putting photos in folders

I’m printing index sheets for the FSPS photos, so that each streets’ group of photos (e.g. Ainslie Road) in the archive folders is divided by a set of A4 colour-printed pages with thumbnails of the photos. These don’t actually have each photo’s URLs or filenames, which I’ve been a bit disappointed about, but it does have the URL of the street’s page. That is enough to get pretty close to an individual photo, and I think it’s good enough. If I were starting this project again I might do things a bit differently, but I’m far enough in now to want to maintain consistency.

I do want to sort out a better URL rewrite for page IDs. At the moment I am including page ID URLs such as https://archives.org.au/Special:Redirect/page/1100 but this would be neater as https://archives.org.au/P1100 (which would prohibit having wiki pages at that URL, but I think that’s okay).

Fremantle Bowling Club

I tried to write about last night at the bowling club, but the WordPress app ate my words without saving a draft. I’d had too much beer (and then mulled wine) for the words to have been any good anyway, so no matter.

I’ve started uploading a few photos to Commons, firstly of the Substation building:

Fremantle Substation 06.jpg
File:Fremantle_Substation_06.jpg (12 July 2019, 16:16:42) by Sam Wilson, CC-BY-SA-4.0.

English: The 1932 electricity substation in Parry Street, Fremantle.

And then of the bowling club.

2019-07-12 1701 Fremantle Bowling Club stage.jpg
File:2019-07-12_1701_Fremantle_Bowling_Club_stage.jpg (12 July 2019, 17:01:04) by Sam Wilson, CC-BY-SA-4.0.

English: The old Fremantle Bowling Club, built c.1950s, shortly before its demolition in 2019.

It’s not a particularly interesting building, and none of the members I spoke to seemed very worried at its demise. The new clubhouse will be bigger and better and generally sounds more exciting.

It was lovely hanging out with the people there, and hearing their stories about the history of the club.

2019-07-12 1726 Fremantle Bowling Club.jpg
File:2019-07-12_1726_Fremantle_Bowling_Club.jpg (12 July 2019, 17:26:04) by User:Samwilson/photography, CC-BY-SA-4.0.

English: Last drinks at the club house: the old Fremantle Bowling Club building, built c.1950s, shortly before its demolition in 2019.

MediaLoader extension

There’s a new MediaWiki extension that’s just been published: MediaLoader. It looks like it’s supposed to load media items such as images, videos, etc. on demand. I haven’t been able to get it to actually work (there’s some strange Composer loading stuff going on in its code) but I think it works by displaying a click-able bit of text such as ‘Load example.jpg’ (not actually a link) that, when clicked, turns into the image or whatever. All it’s doing for me right now is turning into the raw wikitext, but maybe there’s something I’m missing.

I guess the idea is to not download/display the image if its not wanted by the user?

Anyway, it’s new, and it’s always nice to see a new extension being made. Huzza!